Histochem Cell Biol DOI 10.1007/s00418-015-1319-1

ORIGINAL PAPER

Identification of receptor‑type protein tyrosine phosphatase μ as a new marker for osteocytes Karien E. de Rooij1,2 · Martijn van der Velde1 · Edwin de Wilt1 · Martine M. L. Deckers1 · Martineke Bezemer1 · Jan H. Waarsing3 · Ivo Que1 · Alan. B. Chan2 · Eric L. Kaijzel1 · Clemens W. G. M. Löwik1 

Accepted: 24 March 2015 © The Author(s) 2015. This article is published with open access at Springerlink.com

Abstract  Osteocytes are the predominant cells in bone, where they form a cellular network and display important functions in bone homeostasis, phosphate metabolism and mechanical transduction. Several proteins strongly expressed by osteocytes are involved in these processes, e.g., sclerostin, DMP-1, PHEX, FGF23 and MEPE, while others are upregulated during differentiation of osteoblasts into osteocytes, e.g., osteocalcin and E11. The receptortype protein tyrosine phosphatase µ (RPTPμ) has been described to be expressed in cells which display a cellular network, e.g., endothelial and neuronal cells, and is implied in mechanotransduction. In a capillary outgrowth assay using metatarsals derived from RPTPμ-knock-out/ LacZ knock-in mice, we observed that the capillary structures grown out of the metatarsals were stained blue, as expected. Surprisingly, cells within the metatarsal bone tissue were positive for LacZ activity as well, indicating that RPTPμ is also expressed by osteocytes. Subsequent histochemical analysis showed that within bone, RPTPμ is expressed exclusively in early-stage osteocytes. Analysis of bone marrow cell cultures revealed that osteocytes are present in the nodules and an enzymatic assay enabled the quantification of the amount of osteocytes. No apparent bone phenotype was observed when tibiae of RPTPμknock-out/LacZ knock-in mice were analyzed by μCT at

* Karien E. de Rooij [email protected] 1

Experimental Molecular Imaging, Department of Radiology, Leiden University Medical Center, Albinusdreef 2, PO Box 9600, 2300 RC Leiden, The Netherlands

2

Percuros BV, Enschede, The Netherlands

3

Department of Orthopaedics, Erasmus MC, Rotterdam, The Netherlands





several time points during aging, although a significant reduction in cortical bone was observed in RPTPμ-knockout/LacZ knock-in mice at 20 weeks. Changes in trabecular bone were more subtle. Our data show that RPTPμ is a new marker for osteocytes. Keywords  Osteocytes · RPTPμ · Bone · Micro-CT · Animal model

Introduction Osteocytes are the most abundant cell type in bone and are involved in several processes in bone homeostasis and metabolism as extensively reviewed by Bonewald (2011). When osteoblasts become embedded in the newly synthesized bone matrix, they develop dendritic processes, change shape and thus differentiate into osteocytes. During this process, osteocalcin levels are elevated (MikuniTakagaki et al. 1995) and E11 is expressed (Wetterwald et al. 1996; Schulze et al. 1999) which may play a role in osteocyte morphology (Zhang et al. 2006). Osteocytes are involved in phosphate metabolism and matrix mineralization through the proteins PHEX, MEPE, FGF23, BSP and DMP-1, which are all strongly expressed by osteocytes (reviewed in Atkins and Findlay 2012). The osteocytederived protein sclerostin plays a key role in regulation of bone formation by modulating BMP and Wnt signaling (van Bezooijen et al. 2004; Ten Dijke et al. 2008; Moester et al. 2010). The position of the osteocytes within the mineralized matrix and the network of cellular processes make them very well positioned to act as “the nerve cells of the bone.” They have been shown to be sensitive to fluid flow shear stress (Klein-Nulend et al. 1995). Tatsumi et al. (2007) have reported that targeted ablation of osteocytes in

13



mice induced bone loss with deterioration of bone architecture. However, upon mechanical unloading, these mice did not show bone loss compared with control mice. In addition, the absence of connexin 43 in osteoblasts/osteocytes also protected against mechanical unloading, while increasing the effect of mechanical loading (Lloyd et al. 2013; Bivi et al. 2013), implying an important role for osteocytes in mechanical induced bone homeostasis. The receptor-like protein tyrosine phosphatase µ (RPTPµ) belongs to the MAM (meprin/A5/RPTPμ) containing subfamily of transmembrane protein tyrosine phosphatases. RPTPµ is expressed in some types of neuronal cells, lung epithelium, cardiac muscle cells and distinct endothelial cells (Koop et al. 2003). An important common feature of these cell types is the importance of cell– cell interactions. Strikingly, some of these cells also form a cellular network. RPTPµ is involved in cell–cell interactions (Brady-Kalnay and Tonks 1994; Gebbink et al. 1995; Zondag et al. 1995) and may play a role in the formation or maintenance of the cellular network. Accordingly, little RPTPµ expression is found in fenestrated types of endothelium, as present in the liver or the spleen (Bianchi et al. 1999). RPTPµ-knock-out/LacZ knock-in mice, in which the expression of the LacZ gene is controlled by the RPTPµ promoter, allow for the analysis of RPTPµ expression in vivo in various tissues (Koop et al. 2003). These mice were shown to have decreased flow-induced dilatation in mesenteric arteries, implying a role for RPTPµ in mechanotransduction (Koop et al. 2005). To analyze the expression of RPTPμ in endothelial cells in an angiogenesis assay, we have performed capillary outgrowth assays using metatarsals derived from RPTPµ-knock-out/LacZ knock-in mice. To our surprise, not only the outgrowing capillaries were positive for LacZ, but also osteocytes within the metatarsal bone tissue showed LacZ expression. Here, we report that, in bone tissue, RPTPµ is expressed exclusively in the osteocytes and therefore is a new marker for osteocytes.

Materials and methods Mice Homozygous RPTPµ-knock-out/LacZ knock-in mice were obtained from Prof. W.H. Moolenaar (The Dutch Cancer Institute, Amsterdam, The Netherlands). These mice were generated by replacing exon 1 of the RPTPµ gene from the start codon to the splice donor site by the LacZ gene, as described by Koop et al. (2003). Wild-type FVB mice were obtained from Harlan (Horst, The Netherlands). Homozygous RPTPµ-knock-out/LacZ knock-in were crossbred with wild-type FVB mice to generate heterozygous

13

Histochem Cell Biol

RPTPµ-knock-out/LacZ knock-in mice. Subsequently, these heterozygous RPTPµ-knock-out/LacZ knock-in mice were used to breed wild-type and knock-out littermates for μCT analysis of bone architecture. Mice were kept in a controlled 12-h/12-h light/dark cycle and had unlimited access to food and water. All animal experiments were performed with approval of the Ethical Committee for animal experiments of the LUMC. Immunohistochemistry and β‑galactosidase staining RPTPμ-knock-out/LacZ knock-in mice were killed either by decapitation (5-day-old neonatal mice) or by cervical dislocation (adult mice). Long bones and calvariae were prepared and fixed in LacZ fixative (0.1 mM NaH2PO4(H2O)/2 mM MgCl2/5 mM EGTA/0.26 % glutaraldehyde) for 30 min. After washing the bones in LacZ washing buffer (0.1 mM NaH2PO4(H2O)/2 mM MgCl2/0.2 % Nonidet P 40/0.1 % deoxycholate) three times for 10 min, they were stained for β-galactosidase activity in LacZ staining solution (LacZ washing buffer supplemented with 1 mg/ml X-gal, 5 mM K3Fe(CN)6 and 5 mM K4Fe(CN)6(3H2O)) overnight at room temperature. The long bones were postfixed in 3.7 % formaldehyde in PBS overnight at 4 °C, decalcified in 15 % EDTA/0.5 % paraformaldehyde for 2 weeks and embedded in paraffin. Sections (5 µm) were counterstained using the staining procedure for connective tissue according to Von WeichertGiesson. Images were acquired using a color CCD camera mounted on a Nikon Eclipse 610 microscope. The calvariae were postfixed as described, washed twice in PBS and photographed directly using a Contax 167MT camera mounted on a Zeiss microscope. Cultures of metatarsals or bone marrow stromal cells were washed two times in PBS, fixed in LacZ fixative for 3 min, washed in LacZ washing buffer three times for 5 min and stained for β-galactosidase activity in LacZ staining solution overnight at room temperature. Following postfixation in 3.7 % formaldehyde in PBS for 10 min, the cultures were washed twice with PBS and photographed. PECAM-1 staining of metatarsal cultures was performed as described (Deckers et al. 2001). Briefly, the cultures were fixed in zinc macrodex formalin fixative and incubated with ER-MP12 for 16 h at 4 °C. After incubation with a biotinylated secondary antibody, the signal was visualized using the AEC chromogen. Capillary outgrowth assay Capillary outgrowth assays were performed essentially as described (Deckers et al. 2001). In brief, 17-day-old fetuses were isolated from pregnant RPTPμ-knock-out/LacZ knock-in mice, and metatarsals were dissected. Metatarsals

Histochem Cell Biol Table 1  Primers used in quantitative RT-PCR for the analysis of gene expression during osteogenic differentiation of bone marrow-derived stem cells Gene

Forward primer

Reverse primer

β2-Microglobulin

5′-TGACCGGCTTGTATGCTATC-3′

5′-CAGTGTGAGCCAGGATATAG-3′

Alkaline phosphatase

5′-ACACCACAACACGGGCGAGG-3′

5′-TGCCCTCGTTGGCCTTCACG-3′

Osteocalcin

5′ -AGAGACAAGTCCCACACAGCAGC-3′

5′ -TGAAGGCTTTGTCAGACTCAGGGC-3′

LacZ

5′-CTACGTCTGAACGTCGAAAACCCG-3′

5′-GTAGCGGTCGCACAGCGTGTACCAC-3′

Sost

5′-TCCTCCTGAGAACAACCAGAC-3′

5′-TGTCAGGAAGCGGGTGTAGTG-3′

were allowed to attach for 72 h in 175 µl minimal essential medium alpha (α-MEM; Gibco, Carlsbad, CA, USA)/ penicillin/streptomycin (Gibco)/10 % (v/v) heat-inactivated fetal calf serum (FCS; Greiner Bio-One, Kemsmünster, Austria). After the attachment phase, medium was replaced with 250 µl fresh culture medium. Following 7 more days of culture, outgrowth of capillaries was visualized by staining for PECAM-1 or β-galactosidase activity as described above. Bone marrow cell cultures RPTPμ-knock-out/LacZ knock-in mice of 8 weeks old were killed by cervical dislocation. Both femurs and tibiae were isolated, the ends were removed, and the bone marrow was obtained by flushing with 5 ml 10 % FCS in PBS. For one differentiation assay, cells of 2–3 mice were pooled, seeded at a density of 1.5 × 106 cells/well in 12-well plates and cultured in phenol red-free α-MEM (Gibco) supplemented with 10 % (v/v) heat-inactivated FCS, penicillin/ streptomycin, and 50 µg/ml ascorbic acid (BDH Prolabo, VWR International, Radnor, PA, USA). The bone marrow cells were cultured for 21 days, and medium was refreshed every 3–4 days. From day 11 of culture onward, the cultures were either supplemented with 10 mM β-glycerol phosphate (Sigma-Aldrich, St. Louis, MO, USA) alone or stimulated with BMP-4 (50 ng/ml; R&D systems, Minneapolis, MN, USA), BMP-6 (100 ng/ml; R&D systems) or Noggin (250 ng/µl; R&D systems) as well. At day 21, the culture medium was withdrawn and the cell layers were processed for β-galactosidase activity determination either by staining with X-gal or by an enzymatic assay using o-nitro-phenyl-β-D-galactopyranoside (ONPG; SigmaAldrich). Parallel cultures were washed twice with PBS, fixed in 3.7 % formaldehyde in PBS for 5 min and stained with 2 % alizarin red S solution (Sigma-Aldrich).

ZnCl2/0.1 % Triton X-100) overnight at 4 °C. Of this lysate, 100 µl was added to 150 µl of LacZ assay buffer (0.1 M K2HPO4/0.1 % Triton X-100/0.02 M Na2HPO4/2.2 mM KCl/0.2 mM MgSO4/7 mM 2-mercaptoethanol) containing 1 mg/ml o-nitro-phenyl-β-D-galactopyranoside (ONPG) and incubated for 3 h at 37 °C. The amount of o-nitro-phenol formed was determined by measuring the absorption of the assay solution at 405 nm using a ThermoMax microplate reader (Molecular Devices, Sunnyvale, CA, USA). The β-galactosidase activity was corrected for the amount of DNA in the cultures. To release the DNA from the cell debris, the remaining cell lysate was incubated with an equal volume of SCC buffer containing 100 µg/ml proteinase K (Invitrogen, Carlsbad, CA, USA) for 40 h at 56 °C. DNA content was measured using Hoechst 33258 (ICN Biomedicals, Inc., Irvine, CA, USA) and calibrated against a DNA standard (0.5–10 µg/ml herring sperm DNA; Invitrogen). Reverse‑transcribed polymerase chain reaction (RT‑PCR) Total RNA was isolated from bone marrow cell cultures at five time points (days 7, 10, 14, 17 and 21) during differentiation with Trizol LS reagent (Invitrogen). Subsequently, cDNA was synthesized using random primers and M-MLV reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Expression of alkaline phosphatase, osteocalcin, LacZ and Sost was determined by quantitative RT-PCR using the QuantiTect SYBR Green PCR kit (Qiagen, Venlo, The Netherlands) with a CFX96 Touch Real-Time PCR Detection System (BioRad, Hercules, CA, USA). Primers (Eurogentec, Seraing, Belgium) were designed using Primer Blast; sequences are given in Table 1. Measurements were performed in triplicate and analyzed using the 2−ΔΔCt method with β2-microglobulin as internal control (Pfaffl 2001). Micro‑computed tomography

Enzymatic assay for the quantification of β‑galactosidase activity in bone marrow cultures Cell cultures were washed twice with PBS and lysed in 500 µl ALP lysis buffer (10 mM glycine/0.1 mM MgCl2/10 µM

The left tibiae of 8-, 20-, 32-, 44- and 56-week-old RPTPμknock-out/LacZ knock-in mice and wild-type littermates (n  = 5–9 per genotype/age group) were prepared for micro-computed tomography (μCT) by fixing with 3.7 %

13



formaldehyde in PBS for 24 h at 4 °C and stored in 70 % alcohol. They were subsequently scanned using the SkyScan 1076 µCT scanner (SkyScan, Kontich, Belgium) with a source voltage of 40 kV and a current set to 250 µA, using a step size of 0.8° over a trajectory of 180°. Images were made with a voxel size of 9 µm and a frame averaging of 3 to reduce noise. Beam hardening was reduced using a 1-mm aluminum filter. Reconstructions were made using NRecon software (version 1.6.2.0; SkyScan) with a ring artifact correction of 5 and a beam hardening correction set to 20 %. The proximal tibia (0.5–1.5 mm from the growth plate) was chosen as the region of interest to study trabecular and cortical bone. To distinguish calcified tissue from non-calcified tissue, the reconstructed grayscale images were segmented by an automated algorithm using local thresholds (Waarsing et al. 2004). The trabeculae and cortex were separated using PrStackBot-New (software developed by the Orthopedic Research Laboratory, Erasmus MC). The following three-dimensional (3D) bone morphometric parameters were determined using the freely available software package 3D-Calculator (http://www.erasmusmc.nl/47460/386156/ Downloads): cortical volume (Ct.V; mm3), cortical thickness (Ct.Th; µm), endocortical volume (Ec.V; mm3), structure model index (SMI), trabecular separation (Tb.Sp; µm), trabecular volume (Tb.V; mm3), trabecular thickness (Tb. Th; µm) trabecular number (Tb.N; mm−1) and trabecular connectivity (Conn.). The SkyScan CT-Analyzer software was used to determine the following two-dimensional (2D) parameters: cortical area (Ct.Ar; mm2), mean polar moment of inertia (MMI(pol); mm4), cortical perimeter (Ct.Pm; mm) and endocortical perimeter (Ec.Pm; mm). Trabecular bone volume fraction (BV/TV; %) was calculated by dividing trabecular volume by endocortical volume, while trabecular connectivity density (Conn.D; mm−3) was derived by dividing connectivity by endocortical volume. Statistical analysis Values represent mean ± SEM. Statistical analysis of the normally distributed data was performed with Graphpad Prism 5 software (La Jolla, CA, USA) using one-way analysis of variance (ANOVA), followed by the post hoc Bonferroni test. For statistical analysis of µCT data, a Student’s t test was performed. Results were considered significant at p